Document details

Personality assessment based on biosignals during a decision-making task

Author(s): Esteves, Beatriz Gonçalves Crisóstomo

Date: 2017

Persistent ID: http://hdl.handle.net/10362/40258

Origin: Repositório Institucional da UNL

Subject(s): Biosignals; Signal Processing; Feature Selection; Machine Learning; Iowa Gambling Task; Five Factor Model; Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias; Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias; Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias


Description

Due to the emergence of novel acquisition devices and signal processing techniques, the study of electrophysiology and its applications has assumed an important role on the Biomedical Engineering community. Recently, research on this area has expanded to several domains, with the psychophysiology being a proeminent one, more specifically in the field of personality psychology. In this thesis, participants were asked to perform a wildly known decision-making task, the Iowa Gambling Task (IGT), and their biosignals were recorded during this performance with the objective of determining whether changes in biosignals could be related to personality. This project was composed by 71 participants and their biosignals were used to extract meaningful features that together could create a predictive model of personality. For this, all biosignals were processed prior to the feature extraction step and the features were extracted from the entire signals, recorded during the performance of the IGT, and also dividing the task in five blocks. After the extraction, a machine learning algorithm was used to compute the best predictive models for the Five Factor Model (FFM) personality dimensions and for the Maximization and Regret scales, using each biosignal individually and in the end all features from all biosignals. The results showed that the predictive models which use features from all biosignals perform better than the models which use only one biosignal. The Openness to Experience, Agreeableness and Maximization scales are well predicted with features from Electrocardiogram (ECG), the Agreeableness, Maximization and Extraversion scales with Electrodermal Activity (EDA) features and the Extraversion and Openness to Experience scales with features from Blood Volume Pulse (BVP). The hypothesis that personality traits is more expressed in the start of IGT was confirmed since the highest number of features is extracted from the Block 1 of the IGT. The results should be further validated for other populations.

Document Type Master thesis
Language English
Advisor(s) Gamboa, Hugo; Cheetham, Marcus
Contributor(s) Esteves, Beatriz Gonçalves Crisóstomo
facebook logo  linkedin logo  twitter logo 
mendeley logo

Related documents

No related documents